IBR Dynamics and Physics-Informed Learning
IBR Identification and Learned-Model Validation
T1 · Advanced topic · 4 modules
Study excitation, parameter meaning, residual ambiguity, and independent evidence.
Module framework available; full lessons, runnable experiments, and worked practice are in development.
Prerequisites and learning path
The modeling and Neural ODE courses; stability foundations for the validation project.
Modules
-
Excitation, Observation, and Identifiability
Define what the measurements can reveal about selected parameters.
Module outline -
Parameter–Residual Compensation
Inspect equivalent fits and test the assumptions of staged estimation.
Module outline -
Configuration Uncertainty and Fair Baselines
Evaluate models without giving one method hidden information advantages.
Module outline -
Independent Validation of Learned Dynamics
Build evidence that extends beyond fitting the calibration trajectories.
Module outline
Course project
Design an identification study with explicit observations and a fair baseline protocol.
Deliverables
- An observation and excitation contract.
- A baseline, split, noise, and seed protocol.
- A report separating prediction accuracy from parameter recovery.